Stomatal closure is regulated by a complex network of signalling events involving numerous intermediates, among them nitric oxide (NO). Little is known about the signalling events occurring downstream of NO. Previous studies have shown that NO modulates cytosolic calcium concentration and the activation of plasma membrane ion channels. Here we provide evidence that supports the involvement of the lipid second messenger phosphatidic acid (PA) in NO signalling during stomatal closure. PA levels in Vicia faba epidermal peels increased upon NO treatment to maximum levels within 30 min, subsequently decreasing to control levels at 60 min. PA can be generated via phospholipase D (PLD) or via phospholipase C (PLC) in concerted action with diacylglycerol kinase (DGK). Our results showed that NO-induced PA is produced via the activation of both pathways. NO-induced stomatal closure was blocked either when PLC or PLD activity was inhibited. We have shown that PLC- and PLD-derived PA represents a downstream component of NO signalling cascade during stomatal closure.
Plants regulate the uptake of carbon dioxide and water loss through pores, called stomata, located in the epidermis of the aerial parts of the plants. Stomata consist of a pair of specialized cells, named guard cells, which open and close the stomatal pore. Stomatal pore size is regulated by volume variation of the guard cells, caused by osmotic changes, driven by massive ion fluxes that occur mainly through guard cell plasma membrane and tonoplast (MacRobbie 1998). Guard cells sense and rapidly respond to several signals such as light, temperature and humidity, and to hormones such as abscisic acid (ABA), auxin and ethylene (Merritt, Kemper & Tallman 2001; Tanaka et al. 2005; Israelsson et al. 2006). Guard cells have been used for years as a model system for the study of signal transduction processes (Li, Assmann & Albert 2006). Many signalling components are involved in the induction of stomatal closure, among them nitric oxide (NO) has acquired particular interest as a novel signalling component.
NO is a short-lived bioactive gas which is able to diffuse and cross biological membranes (Stamler, Singel & Loscalzo 1992). NO is a well-established second messenger in animals (Davis et al. 2001) and has been shown to be involved in different processes in plants, such as root growth, cell wall lignification, senescence, chlorophyll biosynthesis and stomatal closure (Lamattina et al. 2003). It has been reported that exogenous application of NO increased plant tolerance to drought stress, through the regulation of stomatal movement (Garcia-Mata & Lamattina 2001). It is now known that ABA induces endogenous NO production (Desikan et al. 2002; Garcia-Mata & Lamattina 2002; Neill et al. 2002). NO, in turn, participates in a subset of ABA-evoked responses by increasing cytosolic Ca2+ concentration and by inactivating inward-rectifying K+ channels (K+in) through cGMP/cADPR-dependent processes (Garcia-Mata et al. 2003). In addition, NO has been also reported to block guard-cell outward-rectifying K+ channels (Sokolovski & Blatt 2004). In parallel, the phospholipid second messenger phosphatidic acid (PA) also has been reported to inactivate (K+in) and to induce stomatal closure (Jacob et al. 1999). Moreover, PA levels have been shown to increased in ABA-treated guard cell protoplasts (Jacob et al. 1999). Given that both PA and NO seem to be downstream components of ABA during stomatal movement regulation, we studied a putative relationship between them during stomatal movement regulation.
Lipids or lipid-derived molecules are a group of plant messengers that have been recently described to be involved during biotic and abiotic stress (Meijer & Munnik 2003). Particularly, PA has emerged as a second messenger in plants (Testerink & Munnik 2005). Different pathways can generate PA. Phospholipase C (PLC) hydrolyses the phospholipid phosphatidylinositol 4,5-bisphosphate (PIP2) into IP3, which is able to release Ca2+ from intracellular stores, and diacylglycerol (DAG). In plants, DAG is phosphorylated to PA by DAG kinase (DGK). This pathway is referred as the PLC/DGK pathway (Munnik 2001). Phospholipase D (PLD) hydrolyses structural phospholipids, such as phospatidylcholine (PC), generating PA and choline. In guard cells, PLC inhibition blocks ABA induction of cytosolic Ca2+ oscillations (Staxen et al. 1999) and stomatal closure (Jacob et al. 1999). Among the seven PLC and seven DGK genes present in the Arabidopsis genome, ABA was shown to induce the expression of AtPLC1 (Hirayama et al. 1995). In addition, an increase of PLC1 activity was reported to be required for the induction of ABA-responsive genes (Sanchez & Chua 2001). Furthermore, several studies have demonstrated the involvement of PLD in ABA responses. PLD activity transiently increased upon ABA treatments of guard cell protoplasts (Jacob et al. 1999). Multiple PLD isozymes exist; Arabidopsis has 12, with distinguishable biochemical and regulatory properties (Elias et al. 2002). AtPLDα1 was shown to be present in guard cells, moreover antisense abrogation or gene disruption of PLDα1 impaired stomatal closure induced by ABA (Sang et al. 2001; Zhang et al. 2004). Yet another PLD, AtPLDδ is responsible for PA accumulation in response to dehydration (Katagiri, Takahashi & Shinozaki 2001).
As we stated earlier, both NO and PA have common downstream components involved in the regulation of stomatal closure. In the present work, we combined pharmacological and in vivo biochemical analysis to study the NO-dependent activation of PA signalling in epidermal peels of Vicia faba.
MATERIAL AND METHODS
Broad bean (V. faba) plants were grown in soil:vermiculite (3:1), at 25 °C using a 16 h photoperiod.
Epidermal peel preparation
All experiments were performed with epidermis (epidermal peels) peeled from the abaxial surface of fully expanded leaves of 2- to 4-week-old plants. The peels were pre-incubated in opening buffer [10 mm MES, pH 6.1 (MES titrated to its pKa with KOH) with 10 mm KCl] under white light, to promote stomatal opening (400 µL per well in 24-well plates).
After 3 h pre-incubation in light at 25 °C, chemicals corresponding to the different treatments were added to the opening buffer and subsequently incubated for 1 h. Stomatal apertures were measured from digital pictures made with a Nikon Cooplix 990 (Nikon, Tokyo, Japan) camera coupled to an optical microscope (Nikon Eclipse 2000; Nikon). Then, the pore width was digitally calculated using the image analysis software Matrox Inspector 2.2 (Matrox Electronic System, Dorval, Canada). Aperture values are the mean of 90–120 stomata measured from at least three independent experiments. Values are expressed as mean ± SE of means. Differences between means were statistically analysed using Statistica 6.0 (StatSoft 98) software (StatSoft, Tulsa, OK, USA).
SNAP (S-nitroso-N-acetylpenicilamine; Molecular Probes, Eugene, OR, USA) was prepared as a 10 mM stock solution in 1% DMSO. The PLC inhibitor U73122 (a-(6-(17β)-3-methoxyestra-1,3,5(10)-trien-17yl)amino)(hexyl)-1H-pyrrole-2,5-dione; Sigma, St. Louis, MO, USA) was prepared as a 4 mm stock solution in 100% DMSO. SNP (sodium nitroprusside; Merck, Darmstadt, Germany) and NOC-5 (3-[2-hydroxy-1-(1-methylethyl)-2-nitrosohydrazino]-1-propanamine; Calbiochem, San Diego, CA, USA) were prepared in water. Nitrosoglutathione (GSNO) was prepared according to Stamler & Loscalzo (1992). Chemicals for lipid extraction and silica-60 thin-layer chromatography (TLC) plates were purchased form Merck.
Lipid extraction and quantification
Guard cell phospholipids were labelled by incubating two epidermal peels of 0.25 cm2 (containing an average of approximately 1000 stomata each) for 5 h in 400 µL of opening buffer containing 0.1 µCi 32PiµL−1 (carrier free), under light. Subsequently, epidermal peels were subjected to the different treatments, as indicated. The treatments were stopped by transferring the peels to a 2 mL Eppendorf tube containing a mixture of 170 µL of opening buffer plus 20 µL perchloric acid 50% (v/v). Lipids were extracted by adding 750 µL of CHCl3/MeOH/HCl (50:100:1, v/v) and vortexing for 5 min. The samples were processed as described before (Munnik et al. 1996). Within each experiment, identical amounts of total radioactivity per treatment were loaded in a TLC plate. Finally, lipids were chromatographed using as a mobile phase an ethyl acetate solvent system [the organic upper phase of ethyl acetate/ isooctane/formic acid/water (13:2:3:10, v/v)] or an alkaline solvent system [CHCl3/MeOH/(25%, w/v) NH4OH/H2O (90:70:4:16, v/v)]. Radiolabelled lipids were visualized by autoradiography (BioMax XAR, Kodak, Amsterdam, the Netherlands) and quantified by phosphoimaging (Storm, Molecular Dynamics, Sunnyvale, CA, USA). For each treatment, radioactivity levels of PA were normalized to the amount of radioactivity in the PE and PG spots. Finally, 32PA levels were expressed as a fold increase with respect to the control.
Phospholipid levels of V. faba epidermal peels
Vicia faba epidermal peels were used to study in vivo phospholipid signalling in guard cells. After preparation of epidermal peels, guard cells remain intact and alive, and represent 97.5% of the living cells. This enables us to study phospholipid signalling in guard cells of functional stomata. Previously, in vivo32Pi incorporation into phospholipids was studied by incubating bacteria, yeast, plant tissues, cell suspensions and guard cell protoplasts with 32P orthophosphate. However this method was never used in epidermal peels. Thus, we first analysed the kinetics of 32Pi incorporation into guard cell phospholipids. After different times of incubation with 32Pi, lipids were extracted and separated by TLC using an alkaline solvent as a mobile phase. Figure 1a shows an autoradiograph of a representative time course of 32Pi incorporation into phospholipids of guard cells. The quantification of the different phospholipid species labelled is expressed as 32Pi incorporation at each time point with respect to the level at 300 min (32Pi incorporation, % from total). Figure 1b shows the differences between 32Pi incorporation kinetics in signalling and structural phospholipids. 32Pi was slowly incorporated into structural phospholipids such as phosphatidylcholine (PC), phosphatidylinositol (PI), phosphatidylglycerol (PG) and phosphatidylethanolamine (PE) but quickly incorporated into the signalling phospholipids PA and phosphatidylinositol phosphate (PIP). Differences are due to the rapidly incorporation of 32Pi in the ATP pool which is subsequently used to phosphorylate diacylglycerol (DAG) and PI into PA and PIP, respectively (Munnik 2001). All phospholipid species reached a stable state of labelling within 180–300 min. Therefore, treatment-induced changes in phospholipid were studied after 300 min of labelling.
NO treatment induces the increase of PA in V. faba guard cells
Several signalling components have been described to act downstream of NO during the induction of stomatal closure (Garcia-Mata et al. 2003; Sokolovski & Blatt 2004; Sokolovski et al. 2005). To examine whether NO induces PA signalling, epidermal peels were incubated for 5 h with 32Pi and then treated with the NO donor SNAP for different times. Lipids were analysed by ethyl acetate TLC, because this system allows a better resolution of PA. Figure 2a shows that PA levels in non-treated epidermal peels remain constant throughout the assay. However, the PA level increases significantly within 20 min of application of 200 µm SNAP, the dose commonly used to promote stomatal closure (Garcia-Mata & Lamattina 2001). A maximal 1.9-fold increase is reached at 30 min, which decreases to control levels after 60 min (Fig. 2a,b). Thus, a SNAP dose-response curve was performed at 30 min of treatment. Figure 2c shows that application of 100 µm SNAP induced a 1.2-fold PA accumulation whereas the application of 200 and 400 µm SNAP induced a 1.9-fold accumulation. In addition, PA levels were measured upon treatments with different NO donors. SNP (100 µm), GSNO (200 µm) and NOC-5 (200 µm) induced 1.6-, 1.45- and 1.9-fold increase of PA, respectively (Fig. 2c). These results show that NO increases PA levels in guard cells within 30 min, and that this induction is dose dependent. The fact that different NO donors induced comparable levels of PA indicates that the response depends on the NO released by the donor, and is not due to other properties of the donor molecule itself or other decay levels of by-products (Stamler & Feelisch 1996). This is the first evidence of an NO-dependent increase of PA production in guard cells, suggesting that PA might be acting as a signalling component of NO-induced stomatal closure.
PLD is involved in NO induction of stomatal closure
It has been shown that NO donors promote stomatal closure in epidermal peels of V. faba (Garcia-Mata & Lamattina 2001). Because NO induces PA accumulation in V. faba guard cells, we analysed if the increase of PA was required for NO-induced stomatal closure. PA can be synthesized via two enzymatic pathways, PLD and PLC/DGK. We first studied the effect of blocking the PLD route on SNAP-mediated stomatal closure. We used a primary alcohol (1-butanol) as an inhibitor of the PA production by PLD (for further details, see Munnik et al. 1995; Ritchie & Gilroy 1998). PLDs have the ability to transfer the phosphatidyl group of their substrate to a primary alcohol (transphosphatidylation), which results in the formation of a phosphatidyl alcohol. In the case of 1-butanol this leads to the formation of phosphatidylbutanol (PBut) instead of PA (Munnik et al. 1995). Secondary alcohols such as 2-butanol cannot inhibit PA formation because they are not transphosphatidylation substrates (Munnik et al. 1995). Hence, we measured the stomatal aperture in V. faba epidermal peels incubated with SNAP in the presence or absence of 1- or 2-butanol. Figure 3a shows that, as previously reported (Garcia-Mata & Lamattina 2001), 200 µm SNAP was effective to promote stomatal closure. However, this effect was inhibited in the presence of 0.1% (v/v) 1-butanol. The addition of 2-butanol did not inhibit NO-induced stomatal closure (Fig. 3a). The presence of 1- or 2-butanol did not produce any changes in stomatal aperture of control treatments, ruling out the possibility of any direct effect of the alcohols. These data indicate that PLD-derived PA is necessary for NO-induced stomatal closure.
Because 1-butanol is able to inhibit NO-induced stomatal closure, then SNAP treatments should induce PLD activation. PBut is a marker of in vivo PLD activity that can be visualized in a TLC. Therefore, we measured PBut production in order to confirm the occurrence of PLD activation during NO treatments. With that aim, epidermal peels were incubated in 32Pi for 5 h and then treated with 0.1% (v/v) 1-butanol in the presence or absence of 200 µm SNAP for 30 min. Under these conditions, SNAP induced a 2.6-fold increase of PBut compared to the control treatment (Fig. 3b,c). 1-Butanol leads to the formation of PBut instead of PA. Accordingly, in Fig. 3d we show that the NO-induced PA levels in the presence of 1-butanol are below the level obtained in the absence of 1-butanol (Fig. 2). These results show that NO-induced PLD activation is required for stomatal closure.
PLC is involved in NO induction of stomatal closure
To test whether also PLC is required for NO-induced stomatal closure, experiments using the specific PLC inhibitor U73122 were conducted. Figure 4a shows that 200 µm SNAP induces stomatal closure. Interestingly, the addition of U73122 (0.5 and 1 µm) prevents NO induction of stomatal closure (Fig. 4a). No effect on stomatal aperture values was observed for any of the tested concentrations of U73122 alone. This indicates that also PLC activity is required for NO-dependent induction of stomatal closure. We determined whether PLC contributes to NO-induced PA formation. Epidermal peels pre-incubated for 5 h with 32Pi were treated with 200 µm SNAP or opening buffer (Control), in the presence or absence of 1 µm U73122, for 30 min. Figure 4b,c show that SNAP-induced PA levels were reduced by the PLC inhibitor (from 2.3- to 1.6-fold). These data suggest that PLC is required for NO-induced stomatal closure and that inhibition of PLC affects the NO-induced accumulation of PA.
In this report we describe that (1) NO induces PA accumulation; (2) NO-induced PA formation involves the activation of two enzymes, PLD and PLC; and (3) both enzymatic pathways are required for induction of stomatal closure by NO. The evidence provided indicates the existence of a novel NO-regulated signalling pathway that involves PA production. Here we discuss the signalling of NO and PA and their role in stomatal closure.
Our results show that the NO donor SNAP triggered PA production in V. faba epidermal peels in a time- and dose-dependent manner (Fig. 2). A previous report showed an increase in PA levels upon ABA treatments in guard cell protoplasts (Jacob et al. 1999). Validation of the phospholipid signalling experiment was determined by treating the epidermal peels with ABA (data not shown). The results we obtained are consistent in time and magnitude with ABA-induced PA levels previously reported in guard cell protoplasts. In addition, comparable PA responses were observed upon ABA treatment of other biological systems, albeit using a different phospholipid labelling system (Ritchie & Gilroy 1998; Jacob et al. 1999; Villasuso et al. 2003; Zhang et al. 2004; Katagiri et al. 2005; Zalejski et al. 2005).
Two signalling pathways might generate an NO-dependent induction of PA: the PLD and the PLC/DGK pathway. We used pharmacological and biochemical approaches to study which of the PA synthesis routes are involved in NO-dependent PA generation and stomatal closure. Firstly, we show that 1-butanol, an inhibitor of PA formation by PLD (Fig. 3d) inhibits NO induction of stomatal closure (Fig. 3a). Accordingly, in vivo PLD activation occurs upon NO treatments (Fig. 3b,c). Previous reports have shown (1) the involvement of PA-derived PLD in stomatal closure upon ABA treatments (Jacob et al. 1999; Zhang et al. 2004); (2) a reduced sensitivity of certain PLD mutants to ABA (see further discussion); and (3) the induction of stomatal closure upon exogenous application of PA (Jacob et al. 1999; Zhang et al. 2004). Thus, it might be suggested that NO activates the PLD signalling pathway, generating the PA involved in the induction of stomatal closure. As previously mentioned, a family of PLD isoenzymes with specific biochemical, regulatory and structural properties exists (Elias et al. 2002; Qin & Wang 2002). The Arabidopsis thaliana genome has 12 PLD genes, of which two, PLDα1 and PLDδ, are related to stomatal movements and drought response (Katagiri et al. 2001; Sang et al. 2001; Zhang et al. 2004; Mishra et al. 2006), which suggests these are good candidates to be regulated by NO.
Secondly, we used the specific PLC inhibitor U73122 to study the involvement of PLC in NO-induced PA generation and stomatal closure. Interestingly, PLC inhibition also reduced both PA formation and NO-induced stomatal closure (Fig. 4). These results suggest that PLC-generated DAG might be phosphorylated by DGK, resulting in PA formation. In vivo DGK activity has been measured in different plant systems, using a well-described protocol (for details, see Munnik 2001). Tissue is incubated for a short time with 32Pi, a condition in which PLD substrates (structural phospholipids) will not be labelled. As a consequence, all 32PA that is formed is the result of phosphorylation of DAG by DGK. However, because structural phospholipids are rapidly labelled in V. faba epidermal peels (Fig. 1), in vivo activity of DGK could not be measured using this method. Experiments using DGK inhibitors and DGK mutants might confirm that NO-induced PA is derived from PLC/DGK activity. However, given that DGK is encoded by a multigene family comprised of seven genes, and some isoenzymes have been shown to have different sensitivity to the inhibitors (Gomez-Merino et al. 2005), this will be the subject of future studies.
The other PLC activation product is IP3, which increases cytoplasmic Ca2+ concentrations. Ca2+ can activate some PLDs, all PLCs and some DGKs (Munnik, Irvine & Musgrave 1998). PLC is involved in ABA induction of stomatal closure (Hirayama et al. 1995; Staxen et al. 1999; Hunt et al. 2003), and IP3 has been proposed as the messenger induced during ABA activation of PLC (Lee et al. 1996). There is no reported evidence that PLC contributes to the PA levels required for stomatal closure. Here we showed that PLC activity is required for the NO induction of PA formation during stomatal closure. Thus, our results indicate a role for PA. However, we cannot discard that the observed inhibition of NO-induced PA by U73122 (Fig. 4b,c) could also be explained by a Ca2+-mediated positive regulation of PLD, as proposed for disease resistance in Arabidopsis (Andersson et al. 2006).
There is significant evidence for the involvement of PLD and PLC in drought, hyperosmotic stress, cold response and plant defence (van der Luit et al. 2000; Munnik & Meijer 2001; Ruelland et al. 2002; Yamaguchi et al. 2004, 2005; Andersson et al. 2006). Based on our results we propose the involvement of both enzymes in the NO signal transduction cascade during stomatal closure. An interesting point to be elucidated is the position of these two enzymes in the NO signalling pathway. Two previous reports described the relationship between NO and the phospholipases, albeit in other biological systems. Firstly, NO-stimulated vacuolar H+-ATPase was blocked by 1-butanol in maize seedlings treated with NaCl. This suggests that PLD-derived PA might be involved in NO-regulated H+-ATPase activity (Zhang et al. 2006). However, activation of PLD upon NO treatments was not shown (Zhang et al. 2006). Secondly, in xylanase-elicited tomato cells, NO was critical for inducing PA accumulation via the PLC/DGK pathway (Laxalt et al. 2007). In contrast to what was found in guard cells, tomato cells treated with the NO donor SNAP did not show induction of PLD activity (Laxalt et al. 2007). A plausible explanation is that the PLD isoform activated by NO in guard cells is not present in tomato cell suspensions.
Targets that act downstream of NO-induced PLD and PLC activity can be proposed. For example, the K+ inward-rectifying channels may be targets, as it was reported that PA can inhibit this channel (Jacob et al. 1999). ABI1 protein phosphatase 2C (PP2C) could be another target. PLDα1-derived PA, binds to ABI1 to signal ABA-promoted stomatal closure. Because ABA-insensitive abi1 and abi2 mutants are able to produce NO in response to ABA but are impaired in ABA induction of stomatal closure, it was proposed that NO acts upstream of ABI1 (Desikan et al. 2002, 2004). Conversely, it was reported that during stomatal closure, NO regulates intracellular Ca2+ increased by a cGMP/cADPR-dependent cascade (Garcia-Mata et al. 2003). Alternatively the Ca2+ increase also could be partially due to NO-dependent IP3 production; however, this remains to be shown.
In conclusion, our results suggest that NO activates phospholipid signalling pathways leading to PA production in guard cells and stomatal closure. One of the challenges for future research will be to validate the presented biochemical and pharmacological data by means of genetics in NO-mediated stomatal closure.
We thank A. ten Have, Tasha Teakle, L. Lanteri and M.R. Blatt for critical reading of the manuscript. This work was financially supported by UNMdP (AML, LL), Consejo Nacional de Investigaciones Científicas y Técnicas (CONICET) (AML, CGM, LL), Fundación Antorchas (AML, LL), Third World Academy of Sciences (TWAS) (AML) and Agencia Nacional de Promoción Científica y Tecnológica (ANPCyT) (CGM, LL).